It is good to remember when seeing a study showing a link between blood-type and some outcome: blood-type itself means very little biologically (it is one large

It is good to remember when seeing a study showing a link between blood-type and some outcome: blood-type itself means very little biologically (it is one largely irrelevant gene among tens of thousands), but blood-type is heavily correlated with ancestry/race, which is heavily correlated with income/class, which is heavily correlated with lotsa stuff. When you see a blood-type effect, you are usually seeing a race effect or a class effect. It looks like this study tries to correct for "race [and] ethnicity", but not income. Even when you try to correct for these things, you can often still end up with a residual effect.